Sample-to-sample Pearsons' correlation is calculated from the CPM values of all expressed genes. The resulting heatmap illustrates the pairwise correlation indices, with red indicating higher correlation and blue indicating lower correlation.

Principal Component Analysis (PCA) is a dimensionality reduction algorithm used to summarize the variability structure of an entire dataset into a few latent variables, called principal components. Each component is defined as a linear combination of the original features (for example, genes) and is orthogonal (and thus independent) to all previous components. Components are defined as eigenvector of the covariance matrix of the data, with the first component explaining the largest percentage of variance and subsequent components accounting for progressively smaller proportions of variance.
Multidimensional Scaling (MDS) is also a dimensionality reduction algorithm that seeks to preserve the pairwise distances between samples when projecting data into a lower-dimensional space. Unlike PCA, which identifies linear combinations of features to maximize variance, MDS is more flexible and works directly with a distance matrix, making it suitable for data where the relationships are non-linear or not well captured by variance. If the relationships between the variables in the dataset are primarily linear and the variance structure is well-defined, results from the two algorithms are likely to converge.
In a PCA/MDS scatterplot, each dot represents a sample; dots that are closer together indicate more similar samples, while distant dots suggest highly divergent samples or potential outliers. The coordinates of each dot on the plot are calculated by the algorithms in an unsupervised manner. However, dots can be color-coded based on sample characteristics chosen by the users, enabling the identification of whether clusters of samples are associated with specific experimental or technical variables.
Both analyses were performed on normalized expression values of the top 5000 most variable genes among the expressed ones. The data were preprocessed by centering and scaling to achieve a zero mean and unit variance. The following interactive plots display the distribution of samples in the reduced component spaces and can be used to highlight different sample characteristics as reported in the metadata.

Over-representation and GSEA analysis.

Software Versions lists versions of software tools extracted from file contents.

Suggested text and references to use when describing pipeline usage within the methods section of a publication.URL: https://github.com/nfdata-omics/rnaseq

nextflow run nfdata-omics/rnaseq -r custom_multiqc_test -hub ht_gitlab -params-file nf-params.yml -profile ht_cluster -ansi-log false -resume

• this information is collected when the pipeline is started.URL: https://github.com/nfdata-omics/rnaseq